US8609025B2

US8609025B2 - Single walled carbon nanotubes with functionally adsorbed biopolymers for use as chemical sensors

Info

Publication number: US8609025B2
Application number: US12/373,654
Authority: US
Inventors: Alan T. Johnson, Jr.
Original assignee: University of Pennsylvania Penn
Current assignee: University of Pennsylvania Penn
Priority date: 2005-03-29
Filing date: 2007-07-13
Publication date: 2013-12-17
Also published as: WO2008097261A3; WO2008097261A2; US20100088040A1

Abstract

Chemical field effect sensors comprising nanotube field effect devices having biopolymers such as single stranded DNA or RNA functionally adsorbed to the nanotubes are provided. Also included are arrays comprising the sensors and methods of using the devices to detect volatile compounds.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

This application is the National Stage of International Application No. PCT/US2007/015999, filed Jul. 13, 2007, which is a continuation-in-part of and claims priority to PCT Application No. PCT/US2006/012005, filed Mar. 29, 2006, which claims priority to U.S. Provisional Application Nos. 60/666,232, filed Mar. 29, 2005, and 60/710,708, filed Aug. 22, 2005. This application also claims priority to U.S. Provisional Application No. 60/830,530, filed Jul. 13, 2006. The disclosures of all of the preceding applications are incorporated by reference in their entireties.

GOVERNMENT SUPPORT

Research leading to the disclosed inventions was funded, in part, by the United States Department of Energy, Grant No. DE-FG02-98ER45701, and from the Laboratory for Research on the Structure of Matter, National Science Foundation, Grant No. NSF DMR00-79909. Accordingly, the United States government may have certain rights in the inventions described herein.

FIELD

This invention relates to the field of chemical sensors. The invention also relates to the field of field-effect devices and sensor arrays. The field of the invention also pertains to using sensors both individually, in combination, and in array fashion for detecting compounds.

BACKGROUND

The concept of an “electronic nose” has been an active area of research for some decades. Researchers have been trying to provide such a device e.g., by attempting to couple an array of chemical odor sensors with a pattern-recognition system. NASA, for instance, has been trying to develop such a nose to detect the presence of gases such as ammonia that may be poisonous to astronauts. Ammonia is integral to life on the space station because it carries heat originating inside the station through pipes outside to space. However, ammonia is poisonous and a leak must be detected quickly and stopped. Humans are only capable of detecting ammonia at about 50 ppm, although it can be dangerous at a concentration of only a few parts per million. A suitable sensor, or electronic nose capable of detecting ammonia at such low concentrations is needed.

Considerable interest has been generated by the Department of Homeland Security about the use of electronic devices in detecting volatile compounds to prevent explosive, chemical, or biological attacks. Screening methods developed to detect explosive or toxic chemicals that may be carried through an airport or seaport are, in many ways, a first line of defense in protecting against such attacks. Such methods currently include laboratory analyses of suspected drugs, drug-sniffing dogs, and the ubiquitous X-ray machine. For most laboratory methods to be performed, a great deal of time must be spent in preparing samples for analysis. Small, portable devices would likely reduce the time needed to detect potential threats. It is greatly desired that sensors used for electronic noses and molecular detection exist and function on a very compact—even molecular—scale and exhibit very good electronic properties.

A significant advance in the art would be to provide chemical sensors that are compact and capable of detecting volatile compounds with all-electronic readout. It is desirable that such sensors be capable of easy modification, in order to provide a wider array of potential sensitivities as well as to be able to enhance the sensitivity to known analytes. Similarly, as the sensors currently available suffer from the major drawback of irreversible adsorption, it is desirable to provide sensors that are capable of self-regeneration in order to prolong the usefulness of the device and save on cost. Also needed are methods of detecting such volatile compounds using a device comprising individual sensors or arrays made of compact chemical sensors.

SUMMARY

Electronic noses and other forms of molecular sensors are provided by this invention and are useful in a number of areas. These include determining the freshness of food, identification of spilled chemicals, and diagnosis of disease states through sampling of head spaces of patient odors, sampling of tissues, and body fluids or the like. Other uses are also apparent, including waste water analysis, determination of explosives, evaluation of the contents of containers and vehicles. However, effective, compact, rapid, reliable and controllable sensors are required for these and other sensing purposes. These objectives are attained through the employment of the aspects of the present invention. The wide variability of biopolymers which may form a part of the sensors of the present invention gives rise to huge diversity in the nature of species which may be detected thereby.

The present invention features chemical sensors comprising a substrate, an insulating layer disposed directly adjacent to the substrate, source and drain electrodes disposed directly adjacent to the insulating layer, at least one nanotube disposed directly adjacent to both the source and drain electrodes, and one or more single stranded polynucleotides functionally adsorbed to the nanotube. In some aspects, individual nucleotides such as dNTPs (dATP, dCTP, dGTP, dTTP, dUTP, dITP, and the like) are functionally adsorbed to the nanotube. More than one individual nucleotide can be functionally adsorbed to the nanotube, and preferably, such nucleotides are present in a mixture. The mixture can comprise any combination of nucleotides, and in some preferred embodiments, the combination of nucleotides corresponds to a polynucleotide chain that is known to function successfully in detection of the analyte of interest. Thus, for example, the nucleotides would be present in the mixture at a molar ratio equivalent to their molar ratio in the polymer. The nanotubes used are characterized as being semiconducting in some embodiments. In some embodiments, the substrate is characterized as being semiconducting.

Also featured are chemical sensors adapted to have a gate voltage applied to the substrate and a bias voltage applied between the source and drain electrodes, the nanotube being capable of conducting current between the source and drain electrodes under some conditions of electric field strength local to the nanotube. Disposed directly adjacent to each of the source and drain electrodes is at least one nanotube and one or more single stranded polynucleotides or mixtures of nucleotides functionally adsorbed to the nanotube.

It is preferable that the single stranded polynucleotides is DNA having SEQ ID NO: 3, 4, 5, 6, 7, 8, or 9. In some aspects, the polynucleotide is RNA, and is more preferably RNA having SEQ ID NO: 10 or 11. In some aspects, nucleotides are used. The nanotube with functionally adsorbed polynucleotides or nucleotides interacts with at least one target molecule to affect the electric field local to the nanotube in some exemplary embodiments. Altered current flow in the nanotube accompanies interaction of the molecule to be detected. In many cases, the amount or concentration of target molecule may be quantified through use of the sensors and methods of this invention.

In some embodiments, the chemical sensor further comprises a data processing system, and the data processing system can further comprise a pattern recognition data processing system capable of receiving signal inputs from at least one of the sensors, the data processing system capable of analyzing the received signal inputs to predict the identity of a molecule interacting with the polynucleotide.

The invention also features chemical sensors comprising a substrate, an insulating layer disposed directly adjacent to the substrate, source and drain electrodes disposed directly adjacent to the insulating layer, a first nanotube disposed directly adjacent to both the source and drain electrodes, and at least one additional nanotube disposed directly adjacent to the source and drain electrodes. In some embodiments, a polynucleotide having SEQ ID NO:1 or 2 is functionally adsorbed to the first nanotube, and one or more polynucleotides having SEQ ID NO: 3, 4, 5, 6, 7, 8, 9, 10, or 11, or a mixture of nucleotides, the nucleotides being present in the mixture according to their relative molar ratio in SEQ ID NO:3, 4, 5, 6, 7, 8, 9, 10, or 11 is functionally adsorbed to each of the additional nanotubes.

Also described herein are methods for making a chemical sensor, the chemical sensor comprising a substrate, an insulating layer disposed directly adjacent to the substrate, source and drain electrodes disposed directly adjacent to the insulating layer, and at least one nanotube disposed directly adjacent to both the source and drain electrodes, the method the methods comprising adsorbing a mixture of nucleotides onto the nanotube, the nucleotides being present in the mixture according to their relative molar ratio in SEQ ID NO: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, or 11.

Also featured are methods for identifying one or more molecules, the methods comprising contacting the molecules to one or more polynucleotides functionally adsorbed to a nanotube to give rise to a molecular interaction, the polynucleotides comprising one or more sequences selected from SEQ ID NOs: 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, or 11, the nanotube disposed directly adjacent to source and drain electrodes of a sensing device; the molecular interaction altering the current flow in the nanotube relative to the current flow in the nanotube absent the molecular interaction; and correlating the altered current flow in the nanotube to identify of the molecules. In some aspects, pattern recognition is used to correlate the altered current flow in the nanotube to identify the molecules.

BRIEF DESCRIPTION OF THE DRAWINGS

The summary, as well as the following detailed description, is further understood when read in conjunction with the appended drawings. For the purpose of illustrating the invention, there are shown in the drawings exemplary embodiments of the invention; however, the invention is not limited to the specific methods, compositions, and devices disclosed. In addition, the drawings are not necessarily drawn to scale. In the drawings:

FIG. 1 a is a schematic representation of a chemical sensor showing an insulating layer resting upon a semiconducting substrate, with carbon nanotube upon the semiconducting substrate between source and drain electrodes. ssDNA is shown on top of the carbon nanotube. FIG. 1 b shows a cartoon representation of the sensor, with an exploded view of polynucleotides functionalized to the carbon nanotube.

FIGS. 2 (a) and (b) show AFM images (1 μm×1 μm, z-range 10 nm) and line scans of the same SWNT before (a) and after (b) functionalization with ssDNA. The measured diameter of the bare SWNT is 5.4±0.1 nm, while after application of ssDNA its diameter is 7.2±0.2 nm. The increase in surface roughness in (b) is attributed to non-specific binding of ss-DNA to the SiO₂substrate. (c) Current (I) versus backgate voltage (V_G) characteristic of a bare SWNT-FET sensor (blue), the same device after functionalization with ssDNA sequence 1 and exposed to air (grey), the same ssDNA/SWNT-FET exposed to TMA vapor (black). Source-drain bias voltage is 100 mV.

FIG. 3 shows the change in sensor current upon odor exposure. Currents are normalized to I₀, the value when exposed to air (no odor). (a) Bare SWNT-FET does not respond to methanol vapor (black points). The same device coated with ssDNA having SEQ ID NO:2 shows clear responses to methanol (grey points). (b) A second bare device responds to TMA (black points) but after application of Seq. 2, the response is tripled (grey points). (c) The sensor response to propionic acid (black points) differs in sign and magnitude from the response to methanol (grey points). Light grey data points represent the current baseline (no odor). V_B=100 mV and V_G=0 V for all data sets.

FIG. 4 shows (a) Change in the device current when sarin-simulant DMMP is applied to SWNT-FETs before and after ss-DNA functionalization. (b) Sensor response to DNT.

FIG. 5 shows the final state after 150 ps of a MD simulation that includes ssDNA, a SWNT, and 1000 water molecules (left panel). The ssDNA strand shown has SEQ ID NO:1. The ssDNA has a secondary “pocket” structure that enhances its interaction with analyte molecules. The right panel shows electrostatic potential at the SWNT surface due to ssDNA molecules and cations, where the central region as shown has positive potential, and the end regions (left and right of center) as shown have negative potential. The dominant contribution to the potential in regions directly beneath the ssDNA molecules is from the counterions, making it positive, a counterintuitive result that agrees with experiment.

FIG. 6 shows the response of RNA-functionalized SWNT to exposure to DMMP at 25 ppm. Sensors show no response to water vapor. The response and recovery are rapid. The system returns to the baseline when the DMMP is removed without need for sensor refreshing. Surprisingly, the sign of the response is opposite from that seen for DNA/NT and DMMP (this is also the case for methanol). Most of the responses for RNA/NT are similar to DNA/NT but the RNA/NT tends to be more sensitive (larger response).

DETAILED DESCRIPTION

It has been discovered in accordance with the present invention that not only is the gross composition of a biopolymer important to the sensing of molecular species by functionalized nanosensors, but that the configuration of the biopolymer is important to the sensing. For example, the sequence of a biopolymer such as a nucleic acid has been discovered to be useful in identifying analytes that contact the sensors. Thus, it is now possible for sensors employing oligonucleotides with the same nucleotide base composition, but different nucleotide sequences, to provide different electronic signals when exposed to the same analyte. Thus, the very large degree of variability in even relatively short oligomers of biopolymers permits devices constructed with specific sequences of biopolymers in association with nanostructures across source and drain electrodes on substrates to give rise to very rich information about the molecular species or analyte coming into contact with such biopolymer-nanostructure associations.

Using computational methodologies such as those described and exemplified herein, it is possible to determine electronic signal structures for large numbers of such biopolymer-nanostructure associations that can distinguish molecular species in the environment to a high degree of precision. The minute sizes of sensors of this invention permit the elaboration of highly dense sensor assemblies, assemblies which can signal the presence of molecular species in ways that can be taken together with the signals of other assemblies to determine the identity of the molecular species in detail. Thus, it is possible to generate “fingerprints” for a given molecular species based on electronic signature profiles from various arrays biopolymer-nanostructure associations. For example, sensor arrays having multiple different types biopolymer-nanostructure associations can more accurately identify the presence of a molecular species relative sensors having a single type of biopolymer-nanostructure association. Similarly, sensor arrays comprised of particular biopolymer-nanostructure associations can be used to more accurately identify the presence of a molecular species relative to a different configuration of biopolymer-nanostructure associations.

Applications of the present invention include determining the freshness of food, identification of spilled chemicals, identification of a biowarfare agent, and diagnosis of disease states through sampling of head spaces of patient odors, sampling of tissues, body fluids, and the like. Other uses are also apparent, including waste water analysis, determination of explosives, evaluation of the contents of containers or vehicles. However, effective, compact, rapid, reliable, and controllable sensors are required for these and other sensing purposes. These objectives are attained through the employment of the aspects of the present invention. The wide variability of biopolymers which may form a part of the sensors of the present invention gives rise to hug diversity of the nature of species that can be detected thereby.

The chemical sensors described herein function as high sensitivity gas sensors that exhibit conductivity changes in response to odor application. SWNTs may be arranged into devices that exhibit characteristics similar to all semiconductor field effect transistors (FETs). Such a SWNT-based device may be called a SWNT-FET device. SWNT-FET devices lacking biopolymers on their nanotubes do not exhibit a conductivity response to many volatile gases. In contrast, embodiments of the chemical sensors, or field effect sensors, described herein are responsive to such gases. The responses differ in sign and magnitude for different odors, and the odor response characteristics may depends on the physical characteristics of the biopolymer used, such as the base sequence of a ssDNA polynucleotide. Volatile compounds that do not exhibit a conductivity response in single semiconducting SWNT devices may exhibit conductivity responses in devices with multiple SWNTs.

Without intending to be limited to any particular mechanism or theory of operation, it is believed that the electrical conductivity of a SWNT FET device is sensitive to charged species bound near the nanotube sidewall because such changed species affect the local electric field near the conduction channel formed by the nanotube. However, the nanotube sidewall typically becomes functionalized before SWNT FETs can be useful chemical sensors in accordance with this invention. Functionalization allows a desired molecular species to be preferentially and specifically bound to the nanotube. Functionalization thus enhances the sensitivity of the SWNT detector as well as the potential range of molecules that can be detected by the SWNT-FET. In accordance with the invention, many modes of functionalization may be employed, however, it is preferable to use a biopolymer or biopolymers such as a polynucleotide or polypeptide, although other biopolymers such as carbohydrates can be used. Particularly preferred polynucleotides are capable of adsorbing to the SWNT through the pi-pi stacking effects. In preferred embodiments, the polynucleotides such as ssDNA are reversibly adsorbed to the SWNT to facilitate regeneration of the sensor. Particularly preferred polynucleotides display secondary structures that specifically bind to odor molecules. Most preferred is ssDNA that adsorbs on the SWNT and produces a positive electrostatic potential at the surface of the SWNT.

The chemical sensors described presently can be used as the sensor elements in an array of sensors, each with an individualized response characteristic, coupled to an advanced pattern recognition data processing system. Biopolymers provide an extensive library of compounds for the present sensors. Many are expected to have a binding affinity to SWNT through a pi-pi stacking interaction. Moreover, a variety of methods, such as directed evolution, may be used to create DNA or RNA molecules designed to bind other target molecules, including nucleic acids, proteins, small organic compounds, or even entire organisms. Thus, the chemical sensors of the present invention may be produced with sensitivity to a large variety of compounds.

The present disclosure describes chemical sensors comprising a substrate that may be semiconducting in some embodiments; an insulating layer; source and drain electrodes on the insulating layer; and, in contact with each of the source and drain electrodes, at least one nanotube functionally adsorbed to a biopolymer. In some embodiments, the sensor is adapted to have a gate voltage applied to the substrate and a bias voltage applied between the source and drain electrodes. The nanotube is then capable of conducting current between the source and drain electrodes under some conditions of electric field strength local to the nanotube. The nanotube may be semiconducting.

FIG. 1 depicts a schematic of a chemical sensor that may be used in some embodiments described herein. An insulating layer 22 rests upon a semiconducting substrate 20. The insulating layer 22 may be silicon dioxide or other electrically insulative material known in the art, while the semiconducting layer 20 acts as a back gate. Any semiconducting or conducting material known to those of skill in the art may be used to comprise the back gate. Such materials, including compounds of the materials, may include, but not be limited to, silicon, germanium, gallium, indium, aluminum, gold, copper, diamond, nitride, arsenide and carbide. Source 26 and drain 24 electrodes may be composed of an electrically conductive metal, such as gold, silver, aluminum, chromium, titanium, or copper. The use of an appropriately doped semiconductor material is also contemplated for

electrodes

24 and 26. A carbon nanotube 30 rests between the source 26 and the drain 24 electrodes and is in contact with each. A polynucleotide such as ssDNA 32 is functionalized to the carbon nanotube 30.

Placing biopolymers on the nanotubes of the present chemical sensors gives the device gas-sensing functionality that utilizes individualized binding properties. The binding affinity of biopolymers, such as ssDNA for example, implies that there will be strong binding between the SWNT and the biopolymer. Molecules or volatile compounds bound by the biopolymers on the SWNT will be brought into close contact with the field effect sensor. This affords great compatibility with modern microfabrication techniques, the convenience of electronic readout, small footprint, and ease of fabrication. Useful biopolymers compatible with the present invention include, but are not limited to, polynucleotides such as DNA and RNA, polypeptides, nucleic acid-polypeptide complexes, carbohydrates, aptamers, ribozymes, and all homologs, analogs, conjugates, or derivatives thereof, as well as mixtures thereof. The term “polynucleotide,” generally refers to any polyribonucleotide or polydeoxribonucleotide, which may be unmodified RNA or DNA or modified RNA or DNA. Polynucleotides include, without limitation single- and double-stranded DNA, DNA that is a mixture of single- and double-stranded regions, single- and double-stranded RNA, and RNA that is mixture of single- and double-stranded regions, hybrid molecules comprising DNA and RNA that may be single-stranded or, more typically, double-stranded or a mixture of single- and double-stranded regions. In addition, polynucleotide encompasses triple-stranded regions comprising RNA or DNA or both RNA and DNA. The term polynucleotide also includes DNAs or RNAs containing one or more modified bases and DNAs or RNAs with backbones modified for stability or for other reasons. Modified bases include, for example, tritylated bases and unusual bases such as inosine. A variety of modifications can be made to DNA and RNA; thus, polynucleotide embraces chemically, enzymatically or metabolically modified forms of polynucleotides as typically found in nature, as well as the chemical forms of DNA and RNA characteristic of viruses and cells. Polynucleotide, as used herein, also embraces relatively short nucleic acid chains, often referred to as oligonucleotides. In preferred embodiments of the invention, single stranded DNA is used. These biopolymers provide a comprehensive library of compounds, each with specific binding characteristics.

The biopolymers can be attached by reacting a solution of DNA with other biopolymers, such as RNA, peptides, etc., before application of a derivatized DNA to the nanotube. In other embodiments, a bare chemical sensor (a sensor having no biopolymer on a nanotube) can first be treated with a given DNA oligomer and then a second biopolymer solution applied to the DNA-decorated sensor.

The biopolymers act both as a linker to the SWNT and a binding site for the volatile compound or odor analyte. These two tasks may be performed by using two separate molecules. For example, the ssDNA could act as a linker module for a separate protein molecule that acts as a targeting module. The targeting module in turn is used to attract specific compounds to be measured as suggested by M. Sergi, et. al., (2004) J. Mol. Recog. 17:198-208. Aptamers may be developed for their affinity for specific proteins, and these molecules should retain their designed binding specificity even when associated with a SWNT. Along with the gas phase detection described herein and in the examples below, the chemical sensors of the present invention may be compatible with liquid phase detection.

Nucleic acid hybridization may also be measured using some embodiments. A SWNT forming a part of a sensor of this invention may be decorated with ssDNA, e.g. having the ssDNA applied to the surface of the SWNT. A sample may then be brought into contact with the sensor that contains a nucleic acid or other molecule for which the binding affinity with the superficial DNA is desired to be determined. Operation of the sensor will, in many cases, give rise to information concerning this interaction and the binding affinity may, thus, be determined.

Single stranded DNA decorated SWNTs may be used as electrochemical electrodes. First, a ssDNA strand known or expected to have affinity for an odor analyte of interest is selected. There may be a bundle of SWNTs decorated with ssDNA exposed to a liquid containing the analyte. If the analyte is bound by the SWNT-ssDNA electrode, standard electrochemical measurement (cyclic voltammetry) is expected to show a current flow into the liquid at a voltage that is characteristic of the analyte.

Embodiments of the present invention also include methods of using a chemical sensor comprising a substrate that may be semiconducting in some embodiments; an insulating layer; source and drain electrodes on the insulating layer; and, in contact with each of the source and drain electrodes, at least one nanotube having biopolymer thereupon; said method comprising contacting said sensor with an atmosphere sample and detecting the presence or absence of a volatile compound. Volatile compounds, as used herein, may be considered as compounds vaporized in the atmosphere. The present methods and devices also detect an odor analyte which is considered a sample of the atmosphere to be tested or analyzed.

Chemical sensor arrays are also provided by the present invention. A sensor array comprises a plurality of sensors in accordance with the sensors described above. The arrays may have different sensors or sets of sensors that are sensitive to a different molecule, molecular family, or genus. The arrays also comprise a data processing system. The system may also comprise a pattern recognition system or other control system.

In one embodiment of the invention, pluralities of field-effect devices, or sensors, are employed to discriminate among different target molecules. A number of different sensors are prepared having SWNTs forming a part of them in accordance with this invention, where the SWNTs of the different sensors have different biopolymers thereupon. A single sample of a sensor may comprise an individual field effect device or may comprise dozens or hundreds of identical field effect devices, it being understood that redundancy gives rise to improved accuracy. In any event, a plurality of different sensors is arrayed in a single apparatus with an appropriate controller in electrical communication therewith. Contacting the array with an analyte causes differential responses in the different sensors with the respective currents being evaluated by the controller. The controller is in communication with computational, evaluative, or other display means in order to reflect the interactions of the sensors with the analyte. Through comparison of the results of the analysis either with prior analysis or with expected values, knowledge of the nature of the analyte may be had.

In view of the foregoing, it is apparent that different individual sensors and assemblages or arrays of either the same or different sensors can be used for predefined purposes. Thus, expected responses of sensors in accordance with this invention may be compared with responses from test gasses, liquids, tissues, head spaces, environments samples and other test samples to determine their contents either qualitatively or with identical determination of contents. Quantification may also ensue.

The following examples are provided to describe the invention in greater detail. They are intended to illustrate, not to limit, the invention.

EXAMPLE 1 Preparation of SWNT-FETs

SWNTs were grown by catalytic chemical vapor deposition (CVD) on a SiO₂/Si substrate. FET circuits were fabricated with Cr/Au source and drain electrodes patterned using electron beam lithography and the degenerately doped silicon substrate used as a backgate (FIGS. 1 a and b) (Radosavljevic M et al. (2002) Nano Lett. 2:761-4). For each device, source-drain current I was measured as a function of bias voltage V_Band gate voltage V_Gunder ambient laboratory conditions. Circuits consisting of individual p-type semiconducting nanotubes, where the carriers are positively charged holes, were selected by using only devices that showed a strong decrease in I(V_G) for positive V_G(ON/OFF ratio exceeding 1000).

ssDNA was adsorbed to the SWNT. The ssDNA functionally adsorbed to the SWNT had either of the following sequences, as described in U.S. Patent Application Publication No. 2004/0101851:

5′ GAGTCTGTGGAGGAGGTAGTC 3′,	(SEQ ID NO: 1)
or,

5′ CTTCTGTCTTGATGTTTGTCAAAC 3′.	(SEQ ID NO: 2)

Single stranded oligonucleotides were obtained from Invitrogen (Carlsbad, Calif.) and diluted in distilled water to make a stock solution of 658 μg/ml (SEQ ID NO: 1) or 728 μg/ml (SEQ ID NO: 2). After odor responses of the bare SWNT-FET device were measured, a 500 μm diameter drop of ssDNA solution was applied to the device for 45 min, and then dried under a nitrogen stream. About 25 devices from two different SWNT growth runs were selected for detailed analysis and treated with ssDNA for the experiments.

Statistical analysis of atomic force microscopy (AFM) images of the same tube before and after DNA application showed an increase in the nominal tube diameter from 5.4±0.1 nm to 7.2±0.2 nm, indicating formation of a nanoscale layer of ssDNA on the SWNT surface (FIGS. 2 a, b). For both of the sequences used, application of ssDNA caused the threshold value of V_Gfor measurable conduction to decrease by 3-4 V (FIG. 2 c). This corresponds to a hole density decrease of roughly 400/μm, assuming a backgate capacitance (25 aF/μm) that is typical for this device geometry. Furthermore, the “ON” state conductivity of the ssDNA/SWNT-FET was ˜10% lower than that of the bare device (FIG. 2 c), suggesting weak carrier scattering by the molecular coating.

EXAMPLE 2 ssDNA-SWNT-FET Detection of Propionic Acid, Trimethylamiene, Methanol Dinitrotoluene, and Dimethyl Methylphosphonate

The data below demonstrate that single walled carbon nanotubes coated with single-stranded DNA oligomers show conductance changes when target molecular species are applied to a chemical sensor described herein. The odor responses are different in sign and/or magnitude for different odors and the odor responses also depend on the base sequence of the DNA decorating the SWNTs. SWNTs were fabricated and arrayed for conductance measurements in an FET-like configuration.

Five odors were used with ssDNA-decorated sensor devices of the present invention to yield differential odor responses. These odors are:

Dinitrotoluene (DNT) is of interest because of its similarity to the explosive TNT, although DNT is not explosive. Similarly, dimethyl methylphosphonate (DMMP) is a non-lethal analogue of the nerve agent sarin.

A reservoir of saturated vapor of each Odor was prepared and connected to a peristaltic pump and switching valve array so that the flow of room air directed over the device (0.1 ml/sec) could be electrically diverted to one of the odor reservoirs for a set time, after which the flow reverted to plain air. Odor pulses of 50 sec duration were used. The air or air/analyte mixture was directed towards the sample through a 2±0.1 mm diameter nozzle positioned 6±1 mm above the sample surface. For each analyte, it is estimated that the concentration delivered to the sample was about 3% of the appropriate saturated vapor pressure (Table 1). The source-drain current through the device I_SDwas measured as a function of gate voltage V_Gfor a fixed bias voltage V_B. For each sample (both with and without DNA coating), it was found that V_G=0 was a region of large transconductance (dl/dV_G), where high sensitivity of the SWNT to environmental perturbations is expected. With the gate voltage set to V_G=0V, the current I_SDwas then monitored as the sample (SWNT or SWNT+DNA coating) was exposed to one of the three odors.

TABLE 1

Measured responses of SWNT devices to gaseous analytes.

	vapor	Est.	bare	SWNT + ssDNA	SWNT + ssDNA
	pressure	conc.	SWNT	SEQ ID NO:. 1	SEQ ID NO: 2
Odor	(Torr)	(ppm)	% ΔI/I	% ΔI/I	% ΔI/I

Water	17.5	700	0 ± 1	0 ± 1	0 ± 1
Propionic acid	4	150	0 ± 1	+17 ± 2	+8 ± 1
TMA	500	20000	−9 ± 2	−20 ± 2	−30 ± 2
Methanol	100	4000	0 ± 1	−12 ± 2	−20 ± 2
DMMP	0.6	25	0 ± 1	−14 ± 2	−7 ± 2
DNT	1	40	0 ± 1	−14 ± 4	−4 ± 2

Estimated concentration corresponds to 3% of the saturation vapor pressure. Each quoted sensor response is based on measurements of 5-10 different devices. Uncertainties are the standard deviation of the mean.

Responses of SWNT-FET devices with or without (bare) ssDNA functionalization were compared. It was observed that the bare SWNT (NT+TMA) does not respond to the odor analyte trimethylamine whereas a SWNT adsorbed to an oligonucleotide bearing SEQ ID NO:2 (NT+SEQ ID NO:2+TMA) yields clear responses to the trimethylamine analyte (FIG. 3 b). Similarly, bare SWNT demonstrated minimal sensitivity to methanol (ΔI/I˜1%), whereas SWNT functionalized with ssDNA having SEQ ID NO:2 showed a 20% decrease in transport current (FIG. 3 a). These results indicate that ssDNA increases the binding affinity for methanol or trimethylamine to the device, thereby increasing the sensor response. Even when a bare SWNT sensor responds to a particular gas (ΔI/I=−10% on exposure to TMA, FIG. 3 b), functionalization with ssDNA enhances the molecular affinity and associated response (ΔI/I=−30%). Additional experiments revealed that different odors elicit different current responses from ssDNA-functionalized SWNT-FET sensors. For example, the response to propionic acid of a device functionalized with ssDNA having SEQ ID NO: 1 differs in both sign and magnitude from the response to methanol (FIG. 3 c). The data also demonstrate that a constant sensor response is maintained through multiple odor exposures.

SWNT sensors functionally adsorbed to ssDNA having SEQ ID NO:2 were observed to be sensitive to both DMMP and DNT (FIG. 4 a), whereas sensors functionally adsorbed to ssDNA having SEQ ID NO:1 did not show a response to the DMMP odor analyte (FIG. 4 b). Control experiments verified that the ssDNA-SWNT-FET sensor did not respond to dipropylene glycol, the solvent used for DNT. In addition, no response of the sensor to water vapor, a common background substance, was observed, regardless of whether the sensor was functionally attached to ssDNA having SEQ ID NO:1 or SEQ ID NO:2. The signal-to-noise levels of the experimental measurements depicted in FIG. 4 indicate that detection of volatile compounds with concentrations less than 1 ppm is possible. The DMMP concentration used in this experiment is estimated to be 25 ppm, and the observed response is distinct but modest (ΔI/I ˜−7% for ssDNA SEQ ID NO:1, and −14% for ssDNA SEQ ID NO:2). The results are summarized in Table 2. The results indicate that measured odor responses differ depending on both the sequence of the DNA strand decorating the SWNTs and the identity of the odor being applied.

TABLE 2

Comparison of SWNT response to various odor analytes.

		% ΔI/I	% ΔI/I
	% ΔI/I for	SWNT + ssDNA	SWNT + ssDNA
Gas	SWNT	SEQ ID NO: 1	SEQ ID NO: 2

Water	0	0	0
Propionic	0	+7 ± 2	+10 ± 2
acid
TMA	−15 ± 1	−8 ± 2*	−50 ± 2
Methanol	0	−10 ± 2	−30 ± 2
DMMP	0	0	−10 ± 1
DNT	0	0	−7 ± 2

The data shown are percent change in conductance. Uncertainties are standard deviation of the mean.
(*This sample showed no response to TMA before applying ssDNA having SEQ ID NO: 1).

Additional single stranded DNA oligonucleotides were generated (SEQ ID NOs: 3-9) and used to functionalize SWNT. Following the same experimental protocols described above, these functionalized SWNT were evaluated for their capacity to detect the analytes TMA, methanol, PA, DMMP, and DN, and compared against bare SWNT. Water vapor was used as a standard. The results summarized in Table 3 provide an average of ten separate analyses of each analyte using a separate SWNT sensor. Surprisingly, it was discovered that sensors bearing ssDNA having the same nucleotide base content, but different arrangement in the ssDNA strand yielded different responses to the same analyte. For example, as shown in Table 3, ssDNA bearing SEQ ID NO: 1 and 3 presented different ΔI/I values when exposed to propionic acid, methanol, and DMMP. In addition, ssDNA bearing SEQ ID NO: 2 and 4 presented different ΔI/I values when exposed to TMA, methanol, DMMP, and DNT.

The analytes were also screened against a panel of single and alternating nucleotide polymers, including GT₁₂(SEQ ID NO:5), A₂₁(SEQ ID NO:6), C₂₁(SEQ ID NO:7), G₂₁(SEQ ID NO:8), and T₂₁(SEQ ID NO:9). These sequences were selected to test whether the odor responses of a sensor based on a complex nucleotide sequence (e.g., SEQ ID NO:1) can be understood as a combination of the responses of sensors based on single or alternating nucleotide polymers. The data show that this hypothesis is not correct. The odor responses of a sensor are controlled by the nucleotide sequence as described above. These experiments also demonstrate that single and alternating nucleotide polymers show response to select analyte odors, making them potentially useful in the construction of a sensor array for an electronic nose system. GT₁₂(SEQ ID NO:5) demonstrated a robust response to TMA vapor (Table 3), and also detected methanol. A₂₁(SEQ ID NO:6) demonstrated a strong response to TMA and methanol, and G₂₁(SEQ ID NO:8) showed a robust response to methanol (Table 3).

TABLE 3

Comparison of SWNT response to various odor analytes.

	% ΔI/I	% ΔI/I	% ΔI/I	% ΔI/I	% ΔI/I
	for	SWNT + ssDNA	SWNT + ssDNA	SWNT + ssDNA	SWNT + ssDNA
Gas	SWNT	SEQ ID NO: 1	SEQ ID NO: 2	SEQ ID NO: 3	SEQ ID NO: 4

Water	0 ± 1	0 ± 1	0 ± 1	0 ± 1	0 ± 1
Propionic acid	0 ± 1	17 ± 2	0 ± 1	8 ± 1	10 ± 3
TMA	−9 ± 2	−20 ± 2	N/A	−30 ± 2	14 ± 2
Methanol	0 ± 1	−12 ± 2	−50 ± 2	−20 ± 2	5 ± 1
DMMP	0 ± 1	−14 ± 2	0 ± 1	−7 ± 2	−15 ± 2
DNT	0 ± 1	−14 ± 4	N/A	−4 ± 2	0 ± 2

	% ΔI/I	% ΔI/I	% ΔI/I	% ΔI/I	% ΔI/I
	SWNT + ssDNA	SWNT + ssDNA	SWNT + ssDNA	SWNT + ssDNA	SWNT + ssDNA
Gas	SEQ ID NO: 5	SEQ ID NO: 6	SEQ ID NO: 7	SEQ ID NO: 8	SEQ ID NO: 9

Water	0 ± 1	0 ± 1	0 ± 1	0 ± 1	0 ± 1
Propionic acid	0 ± 1	0 ± 1	0 ± 1	0 ± 1	0 ± 1
TMA	−50 ± 5	−25 ± 3	N/A	N/A	N/A
Methanol	−4 ± 2	−15 ± 2	0 ± 1	−50 ± 2	0 ± 1
DMMP	0 ± 1	0 ± 1	0 ± 1	0 ± 1	0 ± 1
DNT	0 ± 1	0 ± 1	0 ± 1	0 ± 1	±

N/A = combination not tested

Based on the results observed for the various ssDNA oligonucleotides described above, it was hypothesized that RNA could be used to create vapor sensors because 1) RNA should interact with SWNT through an attractive pi-pi interaction in a manner similar to how ssDNA does, 2) it is known that RNA can be engineered for controlled chemical affinity in a manner similar to what is done with ssDNA, and, 3) RNA is generally considered to be a more reactive molecule than ssDNA, so it was anticipated that the resulting vapor sensors would be more sensitive than the corresponding DNA/NT sensors. Accordingly, single stranded oligonucleotides that are RNA analogs of SEQ ID NOs: 1 and 2 were prepared and used to functionalize SWNT. The RNA oligonucleotides are designated SEQ ID NO: 10 and 11. Following the experimental protocols described above, these RNA-functionalized SWNT were evaluated for their capacity to detect the panel of analytes including TMA, methanol, PA, DMMP, and DNT. Water vapor was used as a standard. The results of the RNA-based experiments are summarized below in Table 4. Representative data from RNA-SWNT sensing devices is shown in FIG. 6. The data in the figure show that the RNA sensors share many of the positive characteristics of the DNA/NT sensors, including insensitivity to water vapor, rapid response and recovery, baseline stability, and no need for sensor refreshing. As shown in Table 4, the RNA-functionalized SWNT strongly reacted against all of the analytes; indeed, as hypothesized, the reaction was always larger than the reaction of the corresponding ssDNA functionalized SWNT. Notably, the RNA-sensors reacted differently to the analytes than their DNA counterparts.

TABLE 4

Comparison of RNA-functionalized SWNT
response to various odor analytes.

	ΔI/I (%)	ΔI/I (%)
	SWNT + RNA	SWNT + RNA
Gas	SEQ ID NO: 10	SEQ ID NO: 11

Water	0 ± 1	0 ± 1
Propionic acid	+45 ± 10	+12 ± 2
TMA	−40 ± 10	−23 ± 3
Methanol	+32 ± 5	+20 ± 5
DMMP	+30 ± 5	+35 ± 10
DNT	−20 ± 5	−70 ± 10

In addition to the single stranded nucleic acid chains described above, a defined mixture of non-polymerized dNTPs was prepared, used to functionalize SWNT, and screened for its capacity to detect the panel of analytes, as described above. The mixture was comprised of nucleotides as they are apportioned in the SEQ ID NO:2 polymer, e.g., a molar ratio of 4A:5C:3G:11T. As shown in Table 5 below, the mixture strongly reacted against methanol, to an even greater extent than SEQ ID NOs: 2 or 4, which are comprised of the same nucleotides. These data also support the hypothesis that nucleotide sequence determines sensor responses. In this case, given the very large number of different nucleotide sequences that are possible, it will prove possible to generate large numbers (hundreds or thousands) of sensors with diverse and uncorrelated odor responses as required for an electronic nose system with computational power approaching that of biological olfactory systems.

TABLE 5

Comparison of SWNT + dNTP mixture
response to various odor analytes.

		% ΔI/I SWNT + ssDNA
	Gas	dTNP mixture (S2 Soup)

	Water	0 ± 1
	Propionic acid	0 ± 1
	TMA	N/A
	Methanol	−60 ± 4
	DMMP	0 ± 1
	DNT	N/A

	N/A = Combination not tested.

The number of distinct ssDNA oligonucleotides that can be generated is extremely large. It is expected that an oligonucleotide can readily bind to SWNT via a pi-pi stacking interaction. Accordingly, it is possible to generate a large family of carbon nanotube-based sensors having distinct odor response characteristics, an important building block of “electronic nose” and “electronic tongue” systems.

As a test of response reproducibility, a SWNT device was exposed to 50 cycles of TMA and air exposure (odor and air pulses each 50 seconds in duration). The results demonstrate that the response of the device to the analyte was maintained within 5% even after 50 trials (data not shown). In contrast, other analytical devices are typically irreversibly bound to the analyte, meaning that the sensor must be regenerated for each subsequent analysis. Device-to-device variation in odor response was also observed to be small (Table 1). This excellent reproducibility observed for both a single device and across devices indicates very favorable prospects for quantitative modeling of individual devices and integrated systems.

In a few samples utilizing ssDNA having SEQ ID NO:2, a response to trimethylamine was observed before application of the DNA. This may be the result of some analytes perturbing the electronic structure of the SWNT wall directly. This perturbation may vary sample-to-sample. For these samples, the response was enhanced after applying the DNA solution, for example, a change from −15% to −50% in the transport current can be seen in Table 2. In general, bare SWNT-FET devices did not give odor responses while embodiments of the present invention gave differential odor responses depending on the identity of the odor and the sequence of bases in the ssDNA adsorbed to the SWNT. Note in FIG. 3 that the responses to propionic acid are different in sign and magnitude from the response to TMA.

EXAMPLE 3 Computer Simulations Showing the Possible Mechanism Behind Volatile Sensing Using ssDNA-Functionalized SWNTs

All-atom molecular dynamics (MD) simulations including up to 50,000 water molecules were conducted to investigate adsorption of ssDNA onto SWNTs, and the results compared with observations from experiments utilizing two different ssDNA sequences (SEQ ID NOS: 1 and 2).

Computations were carried out on Tungsten, an Intel Xeon 3.0 GHz Dell Cluster, at the National Center for Supercomputing Applications. These have recently been shifted to a local 128-node cluster. The GROMACS9 software package is employed for the MD. Electrostatic interactions were calculated using the Particle Mesh Ewald10 method. Constant temperature and pressure conditions were imposed using the Berendson thermostat, and Parrinello-Rahman barostat, respectively. (Berendsen H J et al. (1984) J. Chem. Phys. 81:3684-90; and, Parrinello M et al. (1981) J. Appl. Phys. 52:7182-90.) The AMBER99 force field and SPC216 model were used for the ssDNA and water, respectively, due to their reliability and widespread use. An empirical potential for the SWNT was employed that uses a Morse potential for the bonds, a cosine potential for the angles, a Lennard-Jones potential for the van der Waals and steric interactions, and a harmonic potential for dihedral angles. These force fields have been used successfully to study the encapsulation of ssDNA in a SWNT (Gao H et al. (2004) Annu. Rev. Mater. Res. 34:123-50.)

AFM measurements indicate adsorption of a homogenous ssDNA layer on the SWNT with a thickness of approximately 1.8 nm. The observed sequence-dependence of sensor response to gaseous analytes suggests that ssDNA assumed a non-trivial secondary structure when adsorbed on the SWNT. Measurements of device current as a function of backgate voltage implied unambiguously that when ssDNA adsorbs on the SWNT, the electrostatic potential it produces at the surface of the SWNT is positive. Since ssDNA in aqueous solution is known to have a large linear charge that is negative, this last result is strongly counterintuitive. The fact that the maximum device current is decreases by ˜10% implies that carrier scattering due to ssDNA adsorption is weak.

The results from the MD simulations are in agreement with the 3 experimental findings mentioned in the last paragraph. MD simulations (FIG. 5) utilizing ssDNA having SEQ ID NO: 1 adsorbing onto a SWNT showed formation of a molecular layer of 1.5 nm thickness, similar to the observed 1.8 nm. The simulations also revealed non-trivial secondary structure in the ssDNA that may have implications for sensor response to small molecule analytes. Preliminary analysis of the structure suggests it results from a competition between pi-pi stacking of ssDNA bases with the six-member carbon rings that form the nanotube sidewall (“tube stacking”) and with neighboring bases along the ssDNA chain7 (“self stacking”). Intriguingly, it was observed that ssDNA sequences previously identified as superior for SWNT sorting (Zheng M et al., 2003) do not assume such a secondary structure. “Tube stacking” dominates in these sequences, and they lie almost completely flat against the SWNT sidewall (data not shown).

Using the final ssDNA conformation produced by the MD simulation described above, the electrostatic potential due to the combined effect of the ssDNA and its Na+ counterions at the surface of the SWNT was found to be positive, again in agreement with the experiment (FIG. 5). The simulation indicates that the positively charged counterions are not found randomly distributed around the negatively charged phosphate groups of the ssDNA backbone. Instead, they migrate preferentially to regions closer to the SWNT sidewall, and therefore dominate the electrostatic potential in this region.

The present invention is not limited to the embodiments described and exemplified above, but is capable of variation and modification within the scope of the appended claims.

Claims

What is claimed:

1. A chemical sensor comprising a substrate, an insulating layer disposed directly adjacent to the substrate, source and drain electrodes disposed directly adjacent to the insulating layer, at least one nanotube disposed directly adjacent to both the source and drain electrodes, and at least one single stranded polynucleotide having SEQ ID NO:3, 4, 5, 6, 7, 8, 9, 10, or 11 adsorbed to the nanotube.

2. The chemical sensor of claim 1 wherein the substrate is characterized as being semiconducting.

3. The chemical sensor of claim 1 wherein the nanotube is characterized as being semiconducting.

4. The chemical sensor of claim 1 wherein the polynucleotide is capable of interacting with at least one molecule to affect the electric field local to the nanotube.

5. The chemical sensor of claim 4, wherein a molecule interacting with the polynucleotide is capable of giving rise to an altered current flow in the nanotube.

6. A chemical sensor array comprising a plurality of sensors in accordance with claim 1.

7. The chemical sensor array of claim 6 wherein at least one of the sensors is sensitive to a different molecule compared to at least one other sensor.